Skip to contents

Compare validation and test error

val_vs_test.Rd

Plot validation error against test error for a list of models.

Usage

val_vs_test(
  model_list,
  data,
  error_fun,
  regex1 = NULL,
  regex2 = NULL,
  name1 = NULL,
  name2 = NULL,
  legendtitle1 = "spot 1",
  legendtitle2 = "spot 2",
  correlation_label = TRUE,
  file = NULL,
  return_type = c("ggplot", "tibble"),
  plot_theme = ggplot2::theme_minimal(),
  colors = NULL,
  width = 7,
  height = 4,
  quiet = FALSE
)

Arguments

model_list

list of Model objects.

data

Data object. Its cohort attribute must be set. Calculate the test error on the cohort of data.

error_fun

function. Error function to calculate the error between the true and predicted response. For its interface, see val_error_fun_prototype().

regex1

character vector. Regular-expression patterns. For every model, we assign it the group in name1 corresponding to the first pattern its name matches and color it accordingly.

regex2

character vector. Regular-expression patterns. For every model, we assign it the group in name2 corresponding to the first pattern its name matches and shape it accordingly.

name1, name2

character vectors. Names of the groups as defined via regex1 and regex2. Therefore, length(name1) == length(regex1) and length(name2) == length(regex2). If name1 or name2 is NULL, we use the corresponding regex1 or regex2 as the name.

legendtitle1, legendtitle2

string. Legend titles for the grouping according to regex1 and regex2.

correlation_label

logical. Whether to show the correlation coefficient between the validation and test error as a label in the plot.

file

string. File name to save the plot to.

return_type

string. Either "ggplot" or "tibble". See return section for details.

plot_theme

ggplot2 theme. Theme to apply to the plot. If it's NULL, use the ggplot2 default theme.

colors

character vector. Colors used for points.

width, height

numeric. Width and height of the stored plot in inches.

quiet

logical. Whether to suppress messages.

Value

ggplot object if return_type == "ggplot" or the tibble underlying the plot if return_type == "tibble".